Keywords

19.1 Introduction

‘Biomimetic’ and ‘nonbiomimetic’ are important classifications in natural product synthesis. Biomimetic synthesis [1], which takes inspiration from biosynthetic pathways, is a rational and efficient approach to natural product synthesis; this is because (1) the common synthetic intermediates of nature can be used, leading to the diversity-oriented synthesis of a series of natural products, and (2) cheap and easily accessible starting materials can be used. Furthermore, the structures of the natural products are restricted to those accessible by biosynthesis [2,3,4,5]. It is also important to note that the routes of biomimetic syntheses have already been realized in nature, albeit in reaction environments that employ enzymes. Although the structures of natural products have been optimized and refined in nature through evolutionary selection, which can be considered a compound screening in nature [6,7,8], biomimetic total synthesis can also efficiently synthesize structurally novel natural product derivatives by utilizing synthons that are not available in nature.

In contrast, nonbiomimetic total synthesis does not mimic the bond formation patterns or key intermediates of biosynthetic pathways; rather, it seeks more efficient and creative synthetic routes to natural products [9, 10]. There are no restrictions in the design of these synthetic routes, although the feasibility of the routes is completely unknown. Another important point is that nonbiomimetic synthesis can often produce natural product derivatives that have never been synthesized in nature [11, 12].

Our group is engaged in the nonbiomimetic total synthesis of indole alkaloids with a focus on alkyne-based strategies [13,14,15] and photoredox catalysis [16]. Alkynes are extremely important tools in the design of nonbiomimetic syntheses. Although nature can synthesize alkyne-containing natural products, such as acetylenic fatty acids, acetylenic amino acids, enediynes, and bacterial polyynes [9], reactions of alkynes are rarely used in nature; this is illustrated by bio-orthogonal reactions that often rely on alkyne chemistry. Radical-based reductive cyclization onto aromatic rings is another important method in the design of nonbiomimetic synthetic routes because biosynthesis extensively uses oxidative radical reactions, but not their reductive counterparts.

In this chapter, we describe our recent achievements in the nonbiomimetic total synthesis of quinocarcin (1) [17, 18], dictyodendrins A–F (2a–f) [19, 20], and zephycarinatines C and D (3a, 3b) [21] (Fig. 19.1), using alkyne-based approaches or reductive radical spirocyclization. Our efforts in the identification of alkaloid derivatives for drug discovery, which are not accessible by biomimetic approaches, are also described.

Fig. 19.1
Five chemical structures labeled quinocarcin, zephycarinatine C 3 a, zephycarinatine C 3 b, dictyodendrin A and dictyodendrin B, dictyodendrin C, D, F, and G, and dictyodendrin E.

Structures of quinocarcin, dictyodendrins, and zephycarinatines

Full size image

19.2 Total Synthesis of Quinocarcin

The tetrahydroisoquinoline (THIQ) antibiotics constitute a large class of alkaloids (over 60 members) that possess a wide range of structural diversity and biological activities (Fig. 19.2) [22, 23]. In 1983, quinocarcin (1) was isolated from a culture of Streptomyces melanovinaceus by Takahashi and Tomita [24, 25] and was demonstrated to show antiproliferative activity against lymphocytic leukemia. Quinocarcinamide is the oxidized surrogate of quinocarcin, and has a tetracyclic lactam core with a primary alcohol moiety. Tetrazomine and lemonomycin are also members of the quinocarcin family, sharing a 3,8-diazabicyclo[3.2.1]octane core structure. Saframycin A and ecteinascidin 743 are well-known THIQ alkaloids with potent antitumor activity that share the alternative 3,9-diazabicyclo[3.3.1]nonane core structure.

Fig. 19.2
A set of 6 chemical structures labeled quinocarcin, quinocarcinamide, tetrazomine, lemonomycin, saframycin A, and Ecteinascidin 743.

Structures of quinocarcin and related tetrahydroisoquinoline alkaloids

Full size image

A biosynthetic pathway to quinocarcin, based on a nonribosomal peptide synthetase (NRPS) composed of five modules (Qcn12, 13, 15, 17, and 19), was proposed by Oikawa and co-workers in 2013 [26] (Scheme 19.1). Glyoxal derivative 5 is biosynthesized by the Qcn13/12-catalyzed condensation of a long-chain fatty acid with L-alanine, followed by glyceryl unit transfer and reductive cleavage of the resulting thioester. The Pictet-Spengler (PS) reaction of m-tyrosine (4) (derived from phenylalanine) with aldehyde 5, catalyzed by the Qcn17 PS domain, produces thioester 6 with a tetrahydroisoquinoline scaffold, which is reductively converted to bicyclic aldehyde 7. A Mannich reaction of aldehyde 7 with tethered 4,5-dehydroarginine affords pyrrolidine-substituted tetrahydroisoquinoline 9, which can be converted to tetracyclic product 10 via reductive release and N-cyclization. It should be noted that the final transformation of 10 to quinocarcin (1), which may include aminal formation, N- and O-methylation, and oxidation, was not proposed.

Scheme 19.1
A biochemical pathway starts with a compound called m tyrosine and ends with quinocarcin, undergoing various reactions such as Pictet-Spengler, reduction, Mannich, and cyclization.

Proposed biosynthetic pathway to quinocarcin (1)

Full size image

Our retrosynthetic disconnections for quinocarcin (1) are depicted in Scheme 19.2. According to the published procedure by Zhu and Stoltz [27, 28], quinocarcin can be synthesized from the quinocarcinamide derivative 11. This amide would be accessible from 12 via hydrogenation and lactam formation, and compound 12 would be obtained by gold-catalyzed hydroamination [29] of benzylamine derivative 13 bearing an alkynyl pyrrolidine moiety (the first key reaction of this synthesis). The benzylamine derivative 13 could be easily prepared by Sonogashira coupling between phenylglycinol derivative 14 and 2,5-cis-2-alkynylpyrrolidine 15. The pyrrolidine would be stereoselectively synthesized by the base-promoted cyclization of bromoallene 16 (the second key reaction in our synthesis); the 2,5-cis-isomer can be selectively produced from both diastereomers of the bromoallene [30]. Comparing the green bonds in the biosynthetic route (Scheme 19.1) to the red bonds in our synthetic route (Scheme 19.2), it can clearly be seen that our synthetic method does not mimic biosynthesis.

Scheme 19.2
A flowchart has the breakdown of quinocarcin into simpler structures through various chemical reactions like hydrogenation, lactam formation, hydroamination, sonogashira, and bromoallene cyclization.

Retrosynthetic analysis of quinocarcin (1)

Full size image

Our synthesis began with TBDPS-protected γ-butyrolactone 17 (Scheme 19.3). Stereoselective α-propargylation, reductive ring-opening of the resulting 3,5-trans-18 with LiBH4, and acetylation gave protected triol 19 in good overall yield. Introduction of the nitrogen functional group under Mitsunobu conditions and propargylic oxidation with SeO2 afforded 20 as a diastereomeric mixture, which was then converted to bromoallene 16a (dr = 55:45) via mesylation, CuBr·SMe2/LiBr-mediated bromination [31], and removal of the Boc group. As expected, NaH-promoted cyclization of 16a in DMF produced 2,5-cis-pyrrolidine 21 in a highly stereoselective manner (2,5-cis:trans = 96:4; 95% yield) [30]. Finally, 2,5-cis-21 was converted to the pyrrolidine unit (2,5-cis-15), ready for the Sonogashira coupling, via deprotection, oxidation, esterification, and N-methylation.

Scheme 19.3
A complex chemical synthesis process creates a compound named 2, 5 cis 15. It features various molecular structures and reactions, each step marked with specific reagents and conditions, leading to different intermediate compounds.

Synthesis of 2,5-cis-15

Full size image

Next, we proceeded with the asymmetric synthesis of phenylglycinol synthon (R)-14 (Scheme 19.4). We were forced to synthesize the furan derivative 14a to solve the regioselectivity issue in the key hydroamination reaction (vide infra). Regioselective lithiation of 23 with LDA, formylation with DMF, and Wittig reaction of the resulting dihalobenzaldehyde gave styrene derivative 24. Enantioenriched diol 25 (81% ee) was obtained by Sharpless asymmetric dihydroxylation of 24, which was recrystallized from CHCl3 to afford the optically pure diol 25 (> 99% ee). An azido group was introduced into 25 with DPPA under Mitsunobu conditions, and t-BuOK-promoted SNAr reaction of 26 gave dihydrofuran derivative 27. Finally, reduction of the azide and Boc protection gave the phenylglycinol unit (R)-14a in good yield.

Scheme 19.4
A complex chemical synthesis process has various steps, each marked with specific reagents and conditions, leading to the formation of compound R-14 a. The compounds involved are labeled 23, 24, 25, 26, and 27.

Synthesis of (R)-14a

Full size image

With the two building blocks required for the coupling reaction in hand, we proceeded with the total synthesis of quinocarcin (Scheme 19.5). Sonogashira coupling between equimolar amounts of (R)-14a and 2,5-cis-15 provided 13a in 92% yield. After removal of the Boc group, gold-catalyzed hydroamination of 13b in dichloroethane (DCE) successfully produced the desired product through 6-endo-dig cyclization. The resulting unstable enamine was directly converted to tetrahydroisoquinoline derivative 28 (90%, 2 steps) via reduction with NaBH3(CN) in a stereoselective manner. Lactamization of the piperidine with one of the ester groups in 28 was efficiently promoted by heating in acetic acid, forming the quinocarcin core structure. The challenging ring cleavage reaction of dihydrobenzofuran in 11a was achieved via Lewis acid-mediated ring-opening chlorination using BF3·Et2O and SiCl4 in DCE, followed by treatment with CsCl (10 equiv.) in MeCN, to produce phenol derivative 30 in 92% yield [32]. This reaction would proceed through chlorination of the oxazolidinium intermediate 29, formed by treatment of 11a with BF3·Et2O and SiCl4 [33]. Finally, methylation of the phenol with dimethyl sulfate, hydrolysis of the chloromethyl group using AgNO3 in a mixed solvent of acetone/H2O, and hydrolysis and reduction of known intermediate 31 using Stoltz’s procedure [28] successfully produced quinocarcin (1). The spectroscopic data of (–)-quinocarcin we synthesized were consistent with those reported previously.

Scheme 19.5
A chemical synthesis of quinocarcin. 2,5-cis-15 reacts with sodium ascorbate to form compound 13. This undergoes gold-catalyzed hydroamination and reacts with toluene and M e C N to form an intermediate compound 30. Subsequent treatment with liquid nitrogen yields quinocarcin.

Total synthesis of (–)-quinocarcin

Full size image

Next, we would like to describe some of the challenges we experienced in this synthesis and how we serendipitously overcame them. In our model experiments of the hydroamination reaction using simplified substrates 32a, Au(I), as well as a range of other transition-metals such as Cu(I), Pt(II), In(III), and Rh(I), turned out to be ineffective for the desired 6-endo-dig cyclization (Scheme 19.6). Instead, 5-exo-dig cyclization produced 34a as the major product in all cases we examined. One important factor that determines the regioselectivity of this reaction would be the electronic nature of the alkyne; transition-metal complexes increase the cationic character of the alkyne carbon that bears the aryl substituent, thus promoting the 5-exo-cyclization. We then focused on the modification of the substrate by introducing a ring fusion; we expected that the fixing of the nitrogen functional group may change the angle of nucleophilic attack. As expected, the use of the seven-membered acetonide-type substrates 32b and 32c enhanced the 6-endo-dig cyclization, affording 33b in 61% yield and 33c in 31–37%, respectively. Next, to further improve the regioselectivity to withstand the total synthesis, we examined seven-membered ring substrates with a carbonate moiety.

Scheme 19.6
Three chemical reactions detail the outcomes of hydroamination reactions with the transformation of starting compounds to a desired and an undesired product, the catalysts used, and yield percentages.

Model experiments of hydroamination reaction

Full size image

To prepare the seven-membered carbonate 36, we treated diol 35 with triphosgene under basic conditions (Scheme 19.7). Contrary to expectation, we obtained dihydrobenzofuran derivative 36′, presumably through five-membered ring formation from the chlorocarbonate intermediate. Considering that the five-membered ring fusion would strongly promote the desired 6-endo-dig cyclization due to the ring strain of the 6/5/5 ring system in 34d, we proceeded to prepare the hydroamination precursor 32d; this was achieved via the reduction of the azide, Boc protection, and Sonogashira coupling. Fortunately, the gold-catalyzed reaction of 32d using cat. A (5 mol %) in DCE gave the desired six-membered ring 33d in 73% yield as the sole regioisomer. Although we were concerned about how to convert the dihydrobenzofuran to the methyl ether in quinocarcin, we decided to proceed with the total synthesis using a benzofuran-type substrate, taking advantage of the perfect regioselectivity.

Scheme 19.7
A complex chemical synthesis process has various steps, each marked with specific reagents and conditions, leading to the formation of compounds 33 d and 34 d marked with 73% and 0%. Some compounds involved are labeled 35, 36, and 32 d.

Serendipity: unexpected formation of the dihydrofuran derivative and its regioselective hydroamination

Full size image

Next, we prepared benzofuran-type substrate 13a and submitted it to the hydroamination reaction. Our initial attempt using the N-Boc derivative was unsuccessful, resulting in either the recovery of 13a or a low catalyst turnover (Scheme 19.8). We speculated that the steric repulsion between one of the methoxycarbonyl groups and the Boc group might inhibit the hydroamination reaction. Thus, we investigated the reaction using free-amine substrate 13b and obtained the desired product 28b (90%) after NaBH3CN reduction, as described; although, increased catalyst loading was necessary.

Scheme 19.8
A chemical reaction starts with the compound labeled 13 a and b, leading to the formation of compound labeled 28 a prime and b prime. The catalysts involved are cat A U 1 and D C E.

Hydroamination of 13 as a key step of the total synthesis of quinocarcin. a yields after reduction with NaBH3CN

Full size image

As we anticipated, the ring-opening of the dihydrobenzofuran ring was troublesome. In 2004, Zewge reported an efficient C–O bond cleavage of dihydrobenzofuran ring using LiI with SiCl4 and BF3·AcOH (Scheme 19.9) [33]. With these reaction conditions in mind, we expected that the carbonyl oxygen of the adjacent lactam would facilitate the ring-opening of benzofuran under the Lewis acidic conditions to generate oxazolidinium intermediate 29, which could lead to the phenylglycinol derivative. Thus, we treated the dihydrobenzofuran derivative 11a with SiCl4 and BF3·AcOH and observed the formation of a suspension, from which we expected the in-situ generation of oxazolidinium intermediate 29. Unfortunately, our initial attempt at hydrolysis provided only recovered starting material, which can be ascribed to the undesired hydrolytic cleavage of the silyl ether which occurred before the required ring-opening of the oxazolidinium moiety. Reductive treatment of the suspension with Et3SiH, for the preparation of 40, only produced the amino alcohol derivative as the over-reduction product. On the contrary, treatment with t-BuNH2 gave amidine 41 in 52% yield in 2 steps, which strongly suggested the formation of the ring-opening intermediate 29. However, all our efforts to convert 41 to quinocarcin were unsuccessful. After considerable experimentation, we finally found that work-up with excess CsCl gave the chloromethyl derivative 30 in 92% yield [32], which led to the successful total synthesis (Scheme 19.5), as well as the formal synthesis of (–)-quinocarcinamide. This total synthesis was only possible because of the remarkable efforts of Dr. Hiroaki Chiba, whose ‘prepared mind’ allowed him to translate good fortune into success.

Scheme 19.9
A chemical reaction of compound 37 leads to the formation of compound 38 in the presence of S i C l 4, B F 3 dot A C O H, and M e C N. A chemical synthesis process includes compounds 11 a, 29, 39, 40, and 30 leading to compound 41 via reduction, chlorination, and imine formation.

Cleavage of the dihydrobenzofuran ring

Full size image

19.3 Total Synthesis of Dictyodendrins

Dictyodendrins (Fig. 19.1) were isolated from the Japanese marine sponge by the Fusetani group (in 2003) [34] and the Australian marine sponge by the Capon group (in 2012) [35]. They possess broad biological activities, including inhibitory activities toward telomerase and β-site amyloid-cleaving enzyme 1 (BACE1), and are thus potential drug leads for addressing cancer and Alzheimer's disease. Their pyrrolo[2,3-c]carbazole core decorated with various substituents has gathered significant attention from synthetic chemists, and several total syntheses have been reported [36]. However, a diversity-oriented synthetic strategy of dictyodendrins based on the early-stage assembly of the core scaffold, followed by the installation of the substituents, was not reported when we began our study.

In 2017, Ready proposed a biosynthetic pathway to dictyodendrins (Scheme 19.10) [36]. Oxidative coupling of tryptophan (42) and tyrosine (43) affords diketone 44, which is transferred to pyrrole 45 by a Paal-Knorr type condensation with a second molecule of tyrosine. Oxidative decarboxylation of 45 to 46 and the subsequent oxidative aldol-type condensation with a third molecule of tyrosine gives 47. There are two pathways from compound 47 to dictyodendrin A (2a): in the first pathway, oxidative cyclization of 47 to dictyodendrin F (2f) is followed by condensation with another tyrosine molecule to produce dictyodendrin A (2a). In the second pathway, the prior coupling of 47 with tyrosine gives 48, which is transformed into dictyodendrin A (2a) via benzene ring construction. Dictyodendrins A (2a) and F (2f) can be considered key intermediates in the biosynthesis of other dictyodendrins.

Scheme 19.10
A biosynthetic pathway leads to the formation of dictyodendrin. Some compounds involved are tryptophan, tyrosine, 44, 45, 46, and 48.

Proposed biosynthetic pathway to dictyodendrins (2)

Full size image

We envisaged developing an efficient method for the assembly of the dictyodendrin core scaffold followed by the installation of the substituents in a regioselective manner, leading to diversity-oriented synthesis and applications in medicinal chemistry of dictyodendrin derivatives. In 2015, we reported a gold-catalyzed [4 + 2] indole synthesis using conjugated diynes and pyrroles (Scheme 19.11) [37], which proceeds via a double-hydroarylation cascade. We envisaged that this reaction, which efficiently produces 4,7-diarylindoles, could be used to construct the dictyodendrin core structure in combination with nitrene chemistry.

Scheme 19.11
A chemical leads to the formation of indole via hydroarylation and catalysts A u plus and 6 endo dig.

Gold-catalyzed [4 + 2]-type indole synthesis reported by our group

Full size image

Our initial approach to the pyrrolo[2,3-c]carbazole is shown in Scheme 19.12. The gold-catalyzed cyclization of diyne 49 with pyrroles would give 4,7-diarylindole 50 bearing an azido group, and subsequent thermal or transition-metal-catalyzed nitrene insertion would produce pyrrolo[2,3-c]carbazole derivative 52. We anticipated that the control of regioselectivity in the gold-catalyzed [4 + 2] annulation to obtain the desired isomer 50 over 51 would be key to the success of this strategy.

Scheme 19.12
A complex chemical synthesis process of compound 49 reacts with specific reagents and conditions, leading to different intermediate desired and undesired compounds. The processes involved are hydroarylation and arylation. The desired compounds are labeled 50, 52, and 55.

Our initial strategy for the construction of dictyodendrin core and unexpected direct formation of pyrrolo[2,3-c]carbazole 52

Full size image

Thus, Ms. Yuka Matsuda, who developed the gold-catalyzed [4 + 2] indole synthesis [37], investigated the gold-catalyzed reaction between diyne 49 and pyrrole. She unexpectedly found that the reaction directly produced the pyrrolo[2,3-c]carbazoles 52 and 53 as an isomeric mixture. We rationalized this result by the generation of α-iminogold carbene intermediate A, subsequent arylation with pyrrole and intramolecular hydroarylation of the resulting pyrrolylindoles B and C [38]. Although the formation of α-iminogold from simple alkynylanilines and the subsequent reaction with nucleophiles, such as alcohol and anisole, was already reported by Gagosz [39] and Zhang [40], the reaction of diynes with pyrrole as the coupling partner, and the cascade cyclization were not reported. Encouraged by this result, we revised our plan for the total synthesis of dictyodendrins via a gold-catalyzed annulation for the direct formation of pyrrolo[2,3-c]carbazole derivatives.

Our retrosynthetic analysis for the diversity-oriented synthesis of dictyodendrins is shown in Scheme 19.13. Dictyodendrins A–F would be obtained by functional group modification of appropriately substituted pyrrolo[2,3-c]carbazoles 54, 55, and 56. The precursor 54 for dictyodendrin A [41] would be obtained from 55 by the installation of the C2 substituent via acylation with (COCl)2 and Grignard reaction. The intermediate 55 (precursor of dictyodendrins C, D, and F [41]) and 56 (precursor of dictyodendrin E [41]) would be obtained from 57 via bromination and Ullmann coupling with methanol, and addition to p-anisaldehyde where necessary. Sequential functionalization of gold-catalyzed annulation product 52a would afford intermediate 57 through Suzuki–Miyaura coupling and N-alkylation. The conjugated diyne 49a would be prepared by the Cadiot–Chodkiewicz coupling [42] of alkynes 58 and 59. Our synthetic strategy is completely different from the proposed biosynthetic pathway, as can be clearly seen by comparing the green bonds in the biosynthesis (Scheme 19.10) and red bonds in our synthesis (Scheme 19.13).

Scheme 19.13
A complex chemical synthesis process has various molecular structures and reactions leading to different intermediate compounds via processes labeled acylation, Grignard, bromination, and Ullmann, addition to aldehyde, Cadiot-Chodkiewicz, A u cyclization, and N-alkylation.

Retrosynthetic analysis of quinocarcins A–F (2a–f)

Full size image

First, we developed a regioselective synthesis of pyrrolo[2,3-c]carbazole 52a that could withstand the total synthesis (Scheme 19.14). Sonogashira coupling of iodoaniline derivative 60 bearing a tert-butoxy group with trimethylsilylacetylene gave alkyne 61. Desilylation and Cadiot–Chodkiewicz coupling [42] with bromoalkyne 59 gave the conjugated diyne 62, which was transformed into the cyclization precursor 49a via removal of the Boc group and azide formation. After optimizing the gold-catalyzed cyclization, we found that exposure of 49a to N-Boc pyrrole to BrettPhosAu(MeCN)SbF6 (5 mol%) in DCE at 80 °C gave the desired pyrrolo[2,3-c]carbazoles 52a with good regioselectivity (52a:53a = 84:16) [38]. A gram-scale reaction using 49a (2.76 g) and BrettPhosAu(MeCN)SbF6 (162 mg, 2 mol%) also worked well, resulting in the isolation of 52a (2.27 g) in 58% yield.

Scheme 19.14
A complex chemical synthesis process leading to formation of compounds 52 a and 53 a has various molecular structures and reactions leading to different intermediate compounds. Some compounds involved are 60, 61 and 58, 62, 63, and 49 a.

Synthesis of pyrrolo[2,3-c]carbazole 52a

Full size image

With pyrrolocarbazole derivative 52a in hand, we investigated the total and formal synthesis of dictyodendrins C, D, and F, all of which do not possess C2 substituents (Scheme 19.15). Deprotection of the Boc group gave 64 in 92% yield; this compound is the common intermediate for the synthesis of the series of dictyodenrins. Dr. Junpei Matsuoka struggled with the low reactivity of the pyrrolocarbazole derivative, which behaved like a stone, as well as the instability of brominated pyrrolocarbazole derivatives. After many unsuccessful attempts, he found that bromination with N-bromosuccinimide (NBS), alkylation with bromide 65a under aqueous conditions, and Suzuki-coupling with boronic acid 66 worked well for the introduction of the C1 and N3 substituents, producing 57a in 42% yield over 3 steps. Although the installation of the hydroxy group at the C5 position of 57a was also troublesome, we succeeded in dibrominating the compound with NBS, followed by mono-selective debromination using NaBH4 and PdCl2(dppf), to obtain the C5-brominated product 68 in 55% yield. Introduction of the methoxy group via Ullmann coupling between 68 and NaOMe afforded the methoxy derivative 55a, the known precursor of dictyodendrin C, which was converted to dictyodendrin F (2f) by deprotection with BBr3, according to the literature protocol reported by Tokuyama [41]. The total synthesis of dictyodendrin D (2d) was also achieved from 64 using benzyl-protected bromide 65b.

Scheme 19.15
A complex chemical synthesis process leading to the formation of dictyodendrins and compounds 65 a, 65 b, and 66 has various molecular structures and reactions leading to different intermediate compounds. Some compounds involved are labeled 52 a, 64, 65 b, 57 a, 67, 68, and 55 a.

Total and formal syntheses of dictyodendrins C, D, and F

Full size image

Next, we moved on to the total synthesis of dictyodendrin A (2a), which has a (4-hydroxyphenyl)acetate moiety as the C2 substituent (Scheme 19.16). Acylation of the methoxy derivative 55a with oxalyl chloride, followed by treatment with methanol, gave keto-ester 69 (87%). The anisyl group was introduced into 69 by Grignard reaction of the carboxylic acid derived from 69; the resulting ester 54 was obtained in 33% yield (4 steps) after esterification and removal of the hydroxy group. It should be noted that the addition of the Grignard reagent to the keto-ester 69 gave a complex mixture, producing only low yield of the ester 54 (9% after reduction). According to a reported procedure [41], the total synthesis of dictyodendrin A (2a) was accomplished through the removal of the protecting groups.

Scheme 19.16
A complex chemical synthesis process leading to the formation of dictyodendrin A, 2 ahas various molecular structures and reactions leading to different intermediate compounds. Some compounds involved are labeled 55 a, 69, and 54.

Total synthesis of dictyodendrin A

Full size image

The total synthesis of dictyodendrin B is shown in Scheme 19.17. To introduce the acyl group at the C2 position, the lithiation-acylation protocol reported by Fürstner was employed [43]. Thus, mono-selective installation of a bromine atom on 57a using 1.05 equiv. of NBS, lithiation with MeLi/n-BuLi, and nucleophilic addition to anisaldehyde gave alcohol 71. C5-selective bromination with NBS, Ley–Griffith oxidation, and Ullmann coupling of 73 with NaOMe gave ketone 56, a known precursor of dictyodendrin E (2e) [41]. Finally, BCl3-mediated cleavage of the tert-butyl group and functional group modifications [41] gave dictyodendrin B (2b).

Scheme 19.17
A complex chemical synthesis process leading to the formation of dictyodendrin B has various molecular structures and reactions leading to different intermediate compounds. Some compounds involved are labeled 57 a, 70, 71, 72, 73, and 56.

Total and formal syntheses of dictyodendrins B and E

Full size image

We then performed the biological evaluation of the synthesized dictyodendrins and derivatives thereof, which are not accessible by biomimetic synthesis. Because a previous report showed that dictyodendrin F exhibited cytotoxicity to human colon cancer HCT116 cells (IC50 = 27.0 μM) [44], the cytotoxicity of several dictyodendrin analogs toward HCT116 cells was tested. Representative cytotoxicities at 30 μM are shown in Fig. 19.3. Interestingly, the cytotoxicities of the simplified analogs 52a and 64 without C1- and C2-substituents (44–64% cell viability) are comparable to that of dictyodendrin F (55%). However, no cytotoxicity was observed for pyrrolocarbazoles 55a, 56, 57a, and 70–72. Furthermore, some of the simplified analogs of the dictyodendrins displayed CDK2/CycA2 and GSK3β inhibitory activity (data not shown).

Fig. 19.3
Four molecular structures have open and cyclic carbon chains with some groups like N H, N R, O H, M e O A r, O t B u. The structures are labeled 52 a and 64, dictyodendrin F, 55 a, 57 a, and 70, and 56, 71, and 72.

Cytotoxicities (HCT116) after treatment with dictyodendrin derivatives at 30 μM

Full size image

Thus, we have successfully completed the total and formal syntheses of dictyodendrins A–F by using a gold-catalyzed annulation on diynes to construct the pyrrolocarbazole core. The late-stage modification of the core structure allowed for the diversity-oriented synthesis of a series of dictyodendrins. The simplified dictyodendrin analogs showed promising biological activities, exemplifying the utility of nonbiomimetic synthesis in natural product-based drug discovery.

19.4 Total Synthesis of Zephycarinatines

Zephycarinatines (3) were isolated from Zephyranthes carinata Herbert in 2017 by Yao et al. (Scheme 19.18) [45, 46]. These compounds belong to the plicamine-type alkaloids, which are characterized by a unique 6,6-spirocyclic core; the structural analogs zephygranditines (74) and plicamine (75) also belong to this class of alkaloids (75) (Scheme 19.18). The plicamine-type alkaloids and derivatives thereof exhibit a variety of biological activities. For example, zephygranditines (74) display cytotoxicity to cancer cell lines and anti-inflammatory effects by inhibiting NO production in lipopolysaccharide (LPS)-activated macrophages [47]. In contrast, information regarding the bioactivities of zephycarinatines is limited.

Scheme 19.18
A chemical reaction involves a molecular structure with compounds Zephycarinatine C, D, A, and plicamine leading to a biomimetic compound and a nonbiomimetic compound via cyclization and cyclization and oxidation, respectively.

Structures of zephycarinatine and related alkaloids, and synthetic strategy to zephycarinatines

Full size image

A proposed biosynthetic pathway to the plicamine-type alkaloids is shown in Scheme 19.19 [48]. The biosynthesis begins with phenylalanine (76) and tyrosine (43), which undergo several enzymatic processes to yield 4′-O-methylnorbelladine (77). Intramolecular phenol–phenol coupling of 4′-O-methylnorbelladine (77) and 1,4-addition of the amine to the α,β-unsaturated ketone in 78 affords noroxomaritidine (79). Following reduction, oxidation, and methylation, haemanthadine (80) is formed; subsequent ring cleavage of 80 and N-methylation leads to the aldehyde 81, which is then condensed with various amines to provide plicamine-type alkaloids.

Scheme 19.19
A complex chemical synthesis process. The combination of phenylalanine and tyrosine forms the 4 prime-O-methylnorbelladine. The intermediate compounds involved in the process of phenol-phenol coupling and ring cleavage methylation. Some compounds involved are labeled 76, 43, 78, 79, 80, and 81.

Proposed biosynthetic pathway to plicamine-type alkaloids

Full size image

Due to their biological relevance and intriguing structural characteristics, plicamine-type alkaloids have been the focus of numerous synthetic investigations [46]. While there is no report of the total synthesis of zephycarinatines (3), the total synthesis of plicamine (75) has been accomplished [49,50,51]. Plicamine (75) bears a (p-hydroxyphenyl)ethyl group attached to the B-ring nitrogen atom and a methoxy group on the C-ring. Interestingly, its configuration is opposite to that of the zephycarinatines (3). A pivotal step in the synthesis of plicamine-type alkaloids is the construction of the quaternary carbon in the core structure. Previous total synthesis of plicamine (75) employed an intramolecular oxidative coupling of electron-rich aromatics to forge the quaternary carbon center (Scheme 19.18, route a) [49,50,51], mimicking the biosynthetic pathway [48]. In contrast, we conceived a nonbiomimetic strategy for the synthesis of zephycarinatines C (3a) and D (3b). Our approach aims to create the quaternary carbon in a distinct manner from the biosynthetic pathway (route b; nonbiomimetic route) [12], potentially broadening the analog diversity of these compounds. To realize this objective, we devised a reductive radical ipso-cyclization onto the aromatic ring in the presence of a photoredox catalyst, providing straightforward access to the 6,6-spirocyclic core skeleton characteristic of the zephycarinatines (3).

Radical ipso-cyclizations can be categorized as either oxidative or reductive, and both types have attracted attention as powerful tools for the preparation of spirocyclic compounds (Scheme 19.20) [52]. In particular, numerous effective approaches have been developed for oxidative cyclization [52], for instance, the pioneering work by Curran on the ipso-cyclization of aryl radicals onto p-O-aryl-substituted benzamide [53]. In contrast, reductive ipso-cyclization is relatively scarce in the literature. Our group reported a reductive ipso-cyclization mediated by samarium(II), which involves an intramolecular addition of ketyl radicals onto aromatic rings [54, 55]. Yoshimi et al. developed an intramolecular ipso-cyclization that proceeds via the photoinduced decarboxylation of an amino acid analog with an N-(2-phenyl)benzoyl group [56]. More recently, Jui reported a photocatalytic dearomative hydroarylation initiated by the reduction of an aryl halide [57]. However, the application of these methods to total synthesis is still limited, probably due to the difficulty in achieving stereoselective ipso-cyclization of highly functionalized substrates.

Scheme 19.20
A chemical reaction includes processes like oxidative radical ipso-cyclization and reductive radical ipso-cyclization where X = O m e et cetera and X = H et cetera, respectively, in the reactants.

Oxidative and reductive radical ipso-cyclizations

Full size image

To achieve stereoselective ipso-cyclization, we selected the visible-light-mediated decarboxylation reaction of amino acid derivatives. This reaction, which generates carbon radicals under mild conditions using LED irradiation [58, 59], facilitated the total synthesis of plicamine-type alkaloids and their derivatives with various functional groups.

The retrosynthetic analysis of zephycarinatines (3) is depicted in Scheme 19.21. We envisaged incorporating the R1 group in the last stage of the synthesis, thereby enabling the preparation of a diverse range of analogs with different N-substituents. The methoxy group would originate from ketone moiety of 82. Oxidation of the 1,4-diene of 83, followed by 1,4-addition, would allow for the formation of the D ring. The amide 83 was expected to result from the functionalization of the hemiaminal 84. We anticipated that the radical ipso-cyclization of carboxylic acid 85, mediated by visible light, would provide the hemiaminal 84. This process is the pivotal step in the synthesis, necessitating the ipso-cyclization of the α-amino carbon radical intermediate A, derived from the carboxyl radical, in a stereoselective manner. To address this challenge, the oxazolidine substrate 85 was designed with the intension of controlling the chiral center at the α-position, inspired by the self-regeneration of stereocenters (SRS) principle reported by Seebach et al. [60]. The carboxylic acid 85 would derive from carboxylic acid 86 and L-serine 87.

Scheme 19.21
A complex chemical synthesis process leading to a compound with radical intermediate A involves processes like alkylation, reduction and alkoxylation, oxidation, amidation, and ipso-cyclization.

Retrosynthetic analysis of zephycarinatines C and D (3a and b)

Full size image

Initially, the condensation between the known biphenyl-2-carboxylic acid derivative 86 [61] and oxazolidine 88 [62], derived from L-serine (87), was carried out (Scheme 19.22). The addition of mesyl chloride to a mixture of 86, 88, and Et3N provided the amide 89 in 66% yield as a single stereoisomer [63]. The predominant formation of cis-89 could result from the selective acylation of cis-88, arising from the steric difference between the equilibrated cis- and trans-88 via a ring-opening reaction [64]. Subsequently, hydrolysis of the ester 89 was performed with LiOH·H2O to give carboxylic acid 85.

Scheme 19.22
A complex chemical synthesis process leading to the formation of Zephycarinatine C and D has various reactions leading to different intermediate compounds. Some compounds involved are labeled 86, 89, 85, 84, 83, 82, and 90.

Total synthesis of zephycarinatines C and D (3a and b)

Full size image

Next, we investigated the ipso-cyclization of a radical derived from carboxylic acid 85. After optimizing the reaction conditions, we found that treatment of 85 with K2CO3 and photocatalyst [Ir{dF(CF3)ppy}2(dtbpy)]PF6 in MeCN under visible-light irradiation gave the desired product 84 in 58% yield. We then adapted the photochemistry to a continuous flow system to improve the irradiation efficiency of the reaction over the batch process [65]. The optimized condition of the batch process was unsuitable for the flow reaction because K2CO3 exhibits limited solubility in MeCN. Therefore, we investigated reaction conditions that result in a homogeneous mixture. When using a soluble base, such as 1,1,3,3-tetramethylguanidine (TMG), instead of K2CO3, the flow reaction afforded the desired product 84 in 48% yield.

With the requisite spirocyclic core in hand, we went on to investigate the formation of the zephycarinatine D-ring. Hemiaminal 84 underwent transformation into N-methyl amide 83 by the removal of the N,O-acetal group, oxidation of the alcohol with 2-hydroxy-2-azaadamantane (AZADOL) [66], and condensation of the resulting carboxylic acid with methylamine. We then explored the oxidation of 1,4-diene and found that the use of tetrapropylammonium perruthenate (TPAP) and NMO [67] facilitated the oxidation of cyclic 1,4-hexadiene followed by simultaneous 1,4-addition of the N-methyl amide, affording keto derivative 82 in 70% yield.

Next, we turned our attention to the total syntheses of zephycarinatines C and D, which required the conversion of the carbonyl group to the methoxy group. While our initial attempts using NaBH4/CeCl3 and DIBAL-H proved unsuccessful, treatment of 82 with LiAlH4 successfully reduced the carbonyl group, stereoselectively yielding alcohol 90 as the sole diastereomer in 74% yield. We then sought stereoinvertive installation of the C3-methoxy group. In the total synthesis of the plicamine analog obliquine, Ley et al. introduced the methoxy group via the nucleophilic substitution with MeOH at the mesylate derived from the corresponding alcohol and MsCl [68]. Thus, we attempted the mesylation of the alcohol 90 using MsCl and Et3N. However, the desired mesylate was not isolated; instead, an undesired chloride likely formed due to the displacement with chloride originating from MsCl. As an alternative approach, we accomplished the mesylation of 90 with Ms2O, and followed this with MeOH treatment to install the methoxy group with inversion of the stereochemistry, which did not require purification. The following N-alkylation using isopentyl bromide and NaH allowed for completion of the total synthesis of zephycarinatine C (3a) in 23% over 3 steps from 90. When we used MeI as the electrophile, the total synthesis of zephycarinatine D (3b) was achieved in 52% over 3 steps. All spectroscopic data of the synthetic zephycarinatines C and D were in accordance with those documented in the literature [45].

Finally, we assessed the inhibitory effects of zephycarinatine derivatives on NO production using LPS-stimulated RAW264.7 cells [47]. While the natural zephycarinatines C (3a) and D (3b) did not exhibit significant inhibitory activities, synthetic intermediate 82, a keto derivative, demonstrated inhibition of NO production in a dose-dependent manner (IC50 = 65.3 μM).

In summary, we achieved the first total synthesis of zephycarinatines C (2.1% overall yield) and D (4.7% overall yield) from the known acid 86 in 11 steps. The synthesis underscores a nonbiomimetic approach for the stereoselective formation of the B-ring via photocatalytic reductive radical ipso-cyclization. It is worth noting that ketone 82 exhibited a moderate inhibitory effect on LPS-induced NO production. This approach has the potential to broaden the chemical space of plicamine-type alkaloids that are typically not accessible with biomimetic approaches. This total synthesis was accomplished through Ms. Haruka Takeuchi's unwavering dedication.

19.5 Conclusion

We have achieved nonbiomimetic total syntheses of quinocarcin, a series of dictyodendrins, and zephycarinatines C and D, by employing gold-catalyzed cyclization and reductive radical spirocyclization. These synthetic routes, which are very different from the proposed biosynthetic pathways, not only facilitate diversity-oriented synthesis but also exemplify the contribution of synthetic chemistry to drug discovery. Our efforts in nonbiomimetic synthesis will potentially lead to natural product-derived drugs that have never been synthesized in nature.